Portacaval anastomosis (PCA) is the standard treatment for cirrhotic patients who have bled from esophageal varices. Recent controlled investigations have shown that PCA, which is frequently complicated by postshunt hepatic failure and encephalopathy, does not appreciably prolong survival. The hepatic failure-encephalopathy syndrome has been attributed to the total diversion of portal blood from the liver. Warren and Britton have devised distal splenorenal shunts (SRS) that permit portal venous flow to the liver. The present investigation is a collaborative study by experienced groups from the Lemuel Shattuck Hospital in Boston and the West Haven Veterans Administration Hospital to compare in a prospective, controlled trial SRS and PCA in the management of portal hypertension. We plan to evaluate survival, longevity, recurrence of variceal hemorrhage and the incidence of hepatic failure and portalsystemic encephalopathy. All patients will have pre- and postinclusion angiographic and hepatic hemodynamic evaluations and prolonged followup.